1 /* 2 * Copyright (C) 2011-2013 Michael Niedermayer (michaelni@gmx.at) 3 * 4 * This file is part of libswresample 5 * 6 * libswresample is free software; you can redistribute it and/or 7 * modify it under the terms of the GNU Lesser General Public 8 * License as published by the Free Software Foundation; either 9 * version 2.1 of the License, or (at your option) any later version. 10 * 11 * libswresample is distributed in the hope that it will be useful, 12 * but WITHOUT ANY WARRANTY; without even the implied warranty of 13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU 14 * Lesser General Public License for more details. 15 * 16 * You should have received a copy of the GNU Lesser General Public 17 * License along with libswresample; if not, write to the Free Software 18 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA 19 */ 20 21 #ifndef SWR_INTERNAL_H 22 #define SWR_INTERNAL_H 23 24 #include "swresample.h" 25 #include "libavutil/channel_layout.h" 26 27 #include "config.h" 28 29 #define SWR_CH_MAX 32 30 31 #define SQRT3_2 1.22474487139158904909 /* sqrt(3/2) */ 32 33 #define NS_TAPS 20 34 35 #if ARCH_X86_64 36 typedef int64_t integer; 37 #else 38 typedef int integer; 39 #endif 40 41 typedef void (mix_1_1_func_type)(void *out, const void *in, void *coeffp, integer index, integer len); 42 typedef void (mix_2_1_func_type)(void *out, const void *in1, const void *in2, void *coeffp, integer index1, integer index2, integer len); 43 44 typedef void (mix_any_func_type)(uint8_t **out, const uint8_t **in1, void *coeffp, integer len); 45 46 typedef struct AudioData{ 47 uint8_t *ch[SWR_CH_MAX]; ///< samples buffer per channel 48 uint8_t *data; ///< samples buffer 49 int ch_count; ///< number of channels 50 int bps; ///< bytes per sample 51 int count; ///< number of samples 52 int planar; ///< 1 if planar audio, 0 otherwise 53 enum AVSampleFormat fmt; ///< sample format 54 } AudioData; 55 56 struct DitherContext { 57 enum SwrDitherType method; 58 int noise_pos; 59 float scale; 60 float noise_scale; ///< Noise scale 61 int ns_taps; ///< Noise shaping dither taps 62 float ns_scale; ///< Noise shaping dither scale 63 float ns_scale_1; ///< Noise shaping dither scale^-1 64 int ns_pos; ///< Noise shaping dither position 65 float ns_coeffs[NS_TAPS]; ///< Noise shaping filter coefficients 66 float ns_errors[SWR_CH_MAX][2*NS_TAPS]; 67 AudioData noise; ///< noise used for dithering 68 AudioData temp; ///< temporary storage when writing into the input buffer isn't possible 69 int output_sample_bits; ///< the number of used output bits, needed to scale dither correctly 70 }; 71 72 struct SwrContext { 73 const AVClass *av_class; ///< AVClass used for AVOption and av_log() 74 int log_level_offset; ///< logging level offset 75 void *log_ctx; ///< parent logging context 76 enum AVSampleFormat in_sample_fmt; ///< input sample format 77 enum AVSampleFormat int_sample_fmt; ///< internal sample format (AV_SAMPLE_FMT_FLTP or AV_SAMPLE_FMT_S16P) 78 enum AVSampleFormat out_sample_fmt; ///< output sample format 79 int64_t in_ch_layout; ///< input channel layout 80 int64_t out_ch_layout; ///< output channel layout 81 int in_sample_rate; ///< input sample rate 82 int out_sample_rate; ///< output sample rate 83 int flags; ///< miscellaneous flags such as SWR_FLAG_RESAMPLE 84 float slev; ///< surround mixing level 85 float clev; ///< center mixing level 86 float lfe_mix_level; ///< LFE mixing level 87 float rematrix_volume; ///< rematrixing volume coefficient 88 float rematrix_maxval; ///< maximum value for rematrixing output 89 enum AVMatrixEncoding matrix_encoding; /**< matrixed stereo encoding */ 90 const int *channel_map; ///< channel index (or -1 if muted channel) map 91 int used_ch_count; ///< number of used input channels (mapped channel count if channel_map, otherwise in.ch_count) 92 enum SwrEngine engine; 93 94 struct DitherContext dither; 95 96 int filter_size; /**< length of each FIR filter in the resampling filterbank relative to the cutoff frequency */ 97 int phase_shift; /**< log2 of the number of entries in the resampling polyphase filterbank */ 98 int linear_interp; /**< if 1 then the resampling FIR filter will be linearly interpolated */ 99 double cutoff; /**< resampling cutoff frequency (swr: 6dB point; soxr: 0dB point). 1.0 corresponds to half the output sample rate */ 100 enum SwrFilterType filter_type; /**< swr resampling filter type */ 101 int kaiser_beta; /**< swr beta value for Kaiser window (only applicable if filter_type == AV_FILTER_TYPE_KAISER) */ 102 double precision; /**< soxr resampling precision (in bits) */ 103 int cheby; /**< soxr: if 1 then passband rolloff will be none (Chebyshev) & irrational ratio approximation precision will be higher */ 104 105 float min_compensation; ///< swr minimum below which no compensation will happen 106 float min_hard_compensation; ///< swr minimum below which no silence inject / sample drop will happen 107 float soft_compensation_duration; ///< swr duration over which soft compensation is applied 108 float max_soft_compensation; ///< swr maximum soft compensation in seconds over soft_compensation_duration 109 float async; ///< swr simple 1 parameter async, similar to ffmpegs -async 110 int64_t firstpts_in_samples; ///< swr first pts in samples 111 112 int resample_first; ///< 1 if resampling must come first, 0 if rematrixing 113 int rematrix; ///< flag to indicate if rematrixing is needed (basically if input and output layouts mismatch) 114 int rematrix_custom; ///< flag to indicate that a custom matrix has been defined 115 116 AudioData in; ///< input audio data 117 AudioData postin; ///< post-input audio data: used for rematrix/resample 118 AudioData midbuf; ///< intermediate audio data (postin/preout) 119 AudioData preout; ///< pre-output audio data: used for rematrix/resample 120 AudioData out; ///< converted output audio data 121 AudioData in_buffer; ///< cached audio data (convert and resample purpose) 122 AudioData silence; ///< temporary with silence 123 AudioData drop_temp; ///< temporary used to discard output 124 int in_buffer_index; ///< cached buffer position 125 int in_buffer_count; ///< cached buffer length 126 int resample_in_constraint; ///< 1 if the input end was reach before the output end, 0 otherwise 127 int flushed; ///< 1 if data is to be flushed and no further input is expected 128 int64_t outpts; ///< output PTS 129 int64_t firstpts; ///< first PTS 130 int drop_output; ///< number of output samples to drop 131 132 struct AudioConvert *in_convert; ///< input conversion context 133 struct AudioConvert *out_convert; ///< output conversion context 134 struct AudioConvert *full_convert; ///< full conversion context (single conversion for input and output) 135 struct ResampleContext *resample; ///< resampling context 136 struct Resampler const *resampler; ///< resampler virtual function table 137 138 float matrix[SWR_CH_MAX][SWR_CH_MAX]; ///< floating point rematrixing coefficients 139 uint8_t *native_matrix; 140 uint8_t *native_one; 141 uint8_t *native_simd_one; 142 uint8_t *native_simd_matrix; 143 int32_t matrix32[SWR_CH_MAX][SWR_CH_MAX]; ///< 17.15 fixed point rematrixing coefficients 144 uint8_t matrix_ch[SWR_CH_MAX][SWR_CH_MAX+1]; ///< Lists of input channels per output channel that have non zero rematrixing coefficients 145 mix_1_1_func_type *mix_1_1_f; 146 mix_1_1_func_type *mix_1_1_simd; 147 148 mix_2_1_func_type *mix_2_1_f; 149 mix_2_1_func_type *mix_2_1_simd; 150 151 mix_any_func_type *mix_any_f; 152 153 /* TODO: callbacks for ASM optimizations */ 154 }; 155 156 typedef struct ResampleContext * (* resample_init_func)(struct ResampleContext *c, int out_rate, int in_rate, int filter_size, int phase_shift, int linear, 157 double cutoff, enum AVSampleFormat format, enum SwrFilterType filter_type, int kaiser_beta, double precision, int cheby); 158 typedef void (* resample_free_func)(struct ResampleContext **c); 159 typedef int (* multiple_resample_func)(struct ResampleContext *c, AudioData *dst, int dst_size, AudioData *src, int src_size, int *consumed); 160 typedef int (* resample_flush_func)(struct SwrContext *c); 161 typedef int (* set_compensation_func)(struct ResampleContext *c, int sample_delta, int compensation_distance); 162 typedef int64_t (* get_delay_func)(struct SwrContext *s, int64_t base); 163 typedef int (* invert_initial_buffer_func)(struct ResampleContext *c, AudioData *dst, const AudioData *src, int src_size, int *dst_idx, int *dst_count); 164 165 struct Resampler { 166 resample_init_func init; 167 resample_free_func free; 168 multiple_resample_func multiple_resample; 169 resample_flush_func flush; 170 set_compensation_func set_compensation; 171 get_delay_func get_delay; 172 invert_initial_buffer_func invert_initial_buffer; 173 }; 174 175 extern struct Resampler const swri_resampler; 176 177 int swri_realloc_audio(AudioData *a, int count); 178 179 void swri_noise_shaping_int16 (SwrContext *s, AudioData *dsts, const AudioData *srcs, const AudioData *noises, int count); 180 void swri_noise_shaping_int32 (SwrContext *s, AudioData *dsts, const AudioData *srcs, const AudioData *noises, int count); 181 void swri_noise_shaping_float (SwrContext *s, AudioData *dsts, const AudioData *srcs, const AudioData *noises, int count); 182 void swri_noise_shaping_double(SwrContext *s, AudioData *dsts, const AudioData *srcs, const AudioData *noises, int count); 183 184 int swri_rematrix_init(SwrContext *s); 185 void swri_rematrix_free(SwrContext *s); 186 int swri_rematrix(SwrContext *s, AudioData *out, AudioData *in, int len, int mustcopy); 187 void swri_rematrix_init_x86(struct SwrContext *s); 188 189 void swri_get_dither(SwrContext *s, void *dst, int len, unsigned seed, enum AVSampleFormat noise_fmt); 190 int swri_dither_init(SwrContext *s, enum AVSampleFormat out_fmt, enum AVSampleFormat in_fmt); 191 192 void swri_audio_convert_init_aarch64(struct AudioConvert *ac, 193 enum AVSampleFormat out_fmt, 194 enum AVSampleFormat in_fmt, 195 int channels); 196 void swri_audio_convert_init_arm(struct AudioConvert *ac, 197 enum AVSampleFormat out_fmt, 198 enum AVSampleFormat in_fmt, 199 int channels); 200 void swri_audio_convert_init_x86(struct AudioConvert *ac, 201 enum AVSampleFormat out_fmt, 202 enum AVSampleFormat in_fmt, 203 int channels); 204 #endif 205